Introduction

B-cell-mediated antibody responses are essential for eliminating invading pathogens. B-cell receptors (BCRs) expressed on the B-cell surface detect the presence of cognate antigens. The binding of antigen to the BCR leads to the activation of signaling cascades (Reth and Wienands 1997; Dal Porto et al. 2004; Kwak et al. 2019), which induce transcriptional programs that prepare B-cells for proliferation and differentiation (Kurosaki et al. 2010; Shlomchik et al. 2019; Wang et al. 2020). BCR signaling also induces rapid antigen internalization, processing, and presentation for T-cell recognition, which provides the second signal required for B-cell clonal expansion and differentiation to high-affinity antibody-secreting cells and memory B-cells (Song et al. 1995; Gitlin et al. 2014). BCR signaling is tightly regulated by various external and internal factors to activate antibody responses that are specific and also qualitatively and quantitatively matched to the encountered antigen.

Antigen-induced receptor reorganization activates BCRs at the B-cell surface. Antigen binding leads to BCR clustering at lipid rafts, which enables raft-resident kinases, such as the Src kinases Lyn and Fyn, to phosphorylate immunoreceptor tyrosine-based activation motifs (ITAMs) in the cytoplasmic domains of the BCR signaling subunit CD79a/b heterodimer (Reth 1994; Pierce 2002; Sohn et al. 2008). Doubly phosphorylated ITAMs recruit and activate spleen tyrosine kinase (Syk), which in turn activates multiple downstream signaling pathways, including phospholipase Cγ2 (PLCγ2), Bruton’s tyrosine kinase (Btk), Ras-GTPase, and phosphatidylinositol-3 kinase (PI3K), initiating signaling cascades (Kurosaki 2000; Dal Porto et al. 2004; Tanaka and Baba 2020). Signaling activation also induces the recruitment and activation of inhibitory phosphatases, such as SH2-containing tyrosine phosphatase-1 (SHP-1) and phosphatidylinositol-5 phosphatase-1 (SHIP-1), to BCR signaling complexes (Brauweiler et al. 2000; Gross et al. 2009; Franks and Cambier 2018). SHIP-1 and SHP-1 negatively regulate BCR signaling by inactivating the plasma membrane docking lipids for stimulatory kinases, such as Btk and Akt (Aman et al. 1998; Bolland et al. 1998) and dephosphorylating BCR and its downstream signaling molecules (Mizuno et al. 2000; Adachi et al. 2001), respectively. The interplay between the stimulatory kinases and the inhibitory phosphatases controls the balance of antibody responses against pathogens and self. Deficiencies in stimulatory kinases by mutations, such as Btk, cause X-linked agammaglobulinemia (XLA) (Kinnon et al. 1993), while deficiencies in the inhibitory phosphatases SHP-1 or SHIP-1 result in autoimmune diseases (Pao et al. 2007; Leung et al. 2013). However, the mechanisms by which BCR signaling attenuation is initiated and regulated remain elusive.

B-cells encounter both soluble and membrane-associated antigens in vivo. Membrane-associated antigens include antigens on antigen-presenting cells, like follicular dendritic cells in B-cell follicles or germinal centers, and pathogenic cells, like bacteria, parasites, and cancer cells (Batista and Harwood 2009; Depoil et al. 2009; Gonzalez et al. 2009; Cyster 2010). The binding of multi-valent soluble antigen and membrane-associated antigen with any valency induces dynamic reorganization of surface BCRs into microclusters, triggering BCR signaling. Subsequently, surface BCRs continue moving to the plasma membrane region contacting antigen-presenting surfaces or to one B-cell pole in the case of soluble antigen, which leads to the growth and merger of BCR microclusters and the formation of immunological synapses (Carrasco et al. 2004; Harwood and Batista 2010) or supra-molecular activation complexes (Unanue and Karnovsky 1973; Schreiner and Unanue 1977; Tolar et al. 2009b; Harwood and Batista 2010).

Multiple mechanisms have been proposed for signaling initiation by antigen-induced surface BCR reorganization. The binding of surface BCRs to membrane-associated antigen has been shown to induce a conformational change in the BCR extracellular domain, exposing a proximal membrane region of membrane IgM that promotes receptor clustering (Tolar et al. 2009a; Shen et al. 2019). Antigen-binding also changes the conformation of the BCR cytoplasmic domains from a closed to an open form, facilitating the recruitment of signaling molecules (Tolar et al. 2005). Alternatively, antigen-BCR interaction can increase the molecular spacing of BCRs in clusters, facilitating BCR interaction with signaling molecules (Yang and Reth 2010; Kläsener et al. 2014). The inhibitory phosphatases SHP-1 and SHIP-1 are also recruited to BCR clusters (Adachi et al. 2001; Seeley-Fallen et al. 2014). However, the mechanisms underlying phosphatase recruitment and its relationship with BCR signaling activation complexes are unknown.

The actin cytoskeleton is essential for antigen-induced BCR reorganization on the B-cell surface in response to both soluble and membrane-associated antigen (Harwood and Batista 2011; Liu et al. 2013b; Song et al. 2013; Hoogeboom and Tolar 2016). Early signaling of the BCR triggers transient actin depolymerization via Rap GTPase and its downstream target cofilin (Freeman et al. 2011), which disassembles the existing cortical actin network that confines the lateral movement of surface BCRs (Treanor et al. 2010). Following this transient depolymerization, BCR signaling activates rapid actin polymerization through the actin nucleation promoting factors Wiskott-Aldrich Syndrome Protein (WASP) and neuronal-WASP (N-WASP), modulating BCR mobility and leading to F-actin accumulation at BCR-antigen interaction sites (Liu et al. 2013a; Rey-Suarez et al. 2020). Actin polymerization and treadmilling in B-cells stimulated by soluble antigen drive surface BCRs to cluster and move to one pole of the B-cell, forming a BCR cap (Schreiner and Unanue 1977; Liu et al. 2012a). Upon interacting with membrane-associated antigen, B-cells organize actin into two dynamic structures. One treadmills outwards, driving B-cell membrane spreading and expanding the B-cell membrane region contacting the antigen-presenting surface (contact zone), which enables more BCRs to engage antigen (Bolger-Munro et al. 2019). The other creates retrograde flow towards the center of the B-cell contact zone, driving the centripetal movement and growth of BCR microclusters (Liu et al. 2012b). Together, these processes amplify BCR signaling (Batista et al. 2010; Harwood and Batista 2011). The extent of B-cell spreading and the amount of BCR-antigen complexes gathered in the contact zone depend on BCR binding affinity and the density of antigen on the membrane (Fleire et al. 2006). Subsequent to spreading, B-cells contract, reducing the contact area, which drives BCR microclusters to merge into central clusters to form immunological synapses (Fleire et al. 2006; Liu et al. 2013a).

The actin cytoskeleton closely interplays with BCR signaling. We have previously shown that mouse primary B-cells with Btk deficiency (Liu et al. 2011) or double knockouts of the actin nucleation promoting factors WASP and N-WASP (Liu et al. 2013a) fail to spread on antigen-presenting surfaces and establish stable interactions with membrane-associated antigen, drastically reducing BCR signaling. Btk can activate WASP and N-WASP through Vav, the guanine nucleotide exchange factor of Cdc42 and Rac and phosphatidylinositol-5 kinase that generates phosphatidylinositol-4,5-biphosphate (PI4,5P₂) (Sharma et al. 2009; Padrick and Rosen 2010). Surprisingly, B-cell-specific knockout of N-WASP (cNKO) but not WASP germline knockout (WKO) enhances B-cell spreading, delays B-cell contraction, and inhibits the centralization of BCR clusters at the contact zone (Liu et al. 2013a). Consistent with these findings, Bolger-Munro et al. have shown that knockdown of the actin nucleation factor Arp2/3 downstream of WASP and N-WASP disrupts the dynamic actin reorganization induced by membrane-associated antigen required for BCR microcluster growth and merger (Bolger-Munro et al. 2019). Using B-cell-specific and germline knockouts, we showed that in addition to N-WASP, the actin motor non-muscle myosin IIA (NMIIA) (Seeley-Fallen et al. 2022) and the actin-binding adaptor protein Abp1 (Seeley-Fallen et al. 2014) are required for B-cell contraction and BCR central cluster formation. Finally, Wang et al. recently showed that in the presence of the adhesion molecule ICAM on the antigen-presenting surface, B-cells form contractile actomyosin arcs, driving centripetal movement of BCR clusters in the B-cell contact zone (Wang et al. 2022). Together, these findings indicate a critical role of different actin networks in B-cell contraction and BCR signaling.

Our previously studies demonstrate that B-cell contraction is critical for BCR signaling attenuation. Delay and inhibition of B-cell contraction through deficiencies of the actin regulators N-WASP, Abp1, or NMIIA, prolongs and/or increases BCR signaling by enhancing the activation of stimulatory kinases and suppressing the activation of inhibitory phosphatases, which elevates the production of autoantibody levels in mice (Liu et al. 2013a; Seeley-Fallen et al. 2014; Seeley-Fallen et al. 2022). N-WASP and NMIIA are not known to interact with kinases and phosphatases directly. The growth, merger, and centralization of antigen-BCR complexes at the B-cell contact zone during the B-cell contraction phase are associated with the reduced BCR signaling capability (Liu et al. 2013a). These findings suggest that B-cell contraction is one of the mechanisms for inducing BCR signaling attenuation. However, how the actin cytoskeleton facilitates the transition of the spreading to contraction phase and how B-cell contraction switches BCR signaling from amplification to attenuation is unknown.

This study explores the mechanisms by which the actin cytoskeleton remodels from spreading lamellipodia to contractile structures and by which B-cell contraction suppresses BCR signaling. Our results show that the contractile actomyosin structures responsible for B-cell contraction originate from branched actin at spreading lamellipodia. N-WASP/Arp2/3-mediated actin polymerization prolongs the lifetime of the lamellipodial actin structures, enables NMIIA recruitment, and drives their movement to the center of the B-cell contact zone. B-cell contraction increases the molecular density within individual BCR-antigen clusters, which promotes the disassociation of both the stimulatory kinase Syk and the inhibitory phosphatases SHIP-1 from BCR clusters, leading to signal attenuation. Our study reveals a new mechanism underlying BCR signal downregulation.

Results

Arp2/3, activated by N-WASP but not WASP, is required for B-cell contraction

Arp2/3-generated branched F-actin is known to drive lamellipodial expansion for B-cell spreading (Bolger-Munro et al. 2019). However, whether branched F-actin is involved in the subsequent contraction phase is unknown. To address this, we perturbed the polymerization of branched F-actin using the Arp2/3 inhibitor CK-666 (50 µM) while using its inactive derivative CK-689 as a control. Pre-warmed splenic B-cells from WT C57BL/6 mice were incubated with monobiotinylated Fab’ fragment of anti-mouse IgM+G attached to planar lipid bilayers (Fab’-PLB) by biotin-streptavidin interaction and imaged live at 37°C using interference reflection microscopy (IRM). The contact area of B-cells treated with CK-689 rapidly increased upon contacting Fab’-PLB and reached a maximum at ≥0.5 min after the initial contact (Figure 1A, Figure 1-figure supplement 1, and Figure 1-Video 1A). Following maximal spreading, most B-cells reduced the area of their contact zone, indicating contraction (Figure 1A, Figure 1-figure supplement 1, and Figure 1-Video 1A). We classified a B-cell as contracting if its contact zone area reduced by ≥5% for at least 10 sec after reaching a maximum value. Based on the average timing for B-cell maximal spreading, we treated B-cells with CK-666 at the beginning of the incubation (Time 0), before B-cell spreading initiation, or at 2 min, when all B-cells had already spread. The effectiveness of CK-666 was detected by reduced Arp2/3 staining in the contact zone (Figure 1-figure supplement 2). As expected, CK-666 treatment at 0 but not 2 min reduced the kinetics of B-cell spreading (Figure 1-figure supplement 3). Importantly, CK-666 treatment at 0 and 2 min both significantly reduced the percentage of B-cells undergoing contraction (Figure 1A-C and Figure 1-Video 1A). The inhibitory effect of CK-666 on B-cell contraction, particularly CK-666 treatment after B-cell spreading, suggests that Arp2/3-mediated branched actin polymerization is required for B-cell contraction.

Figure 1.

WASP and N-WASP are actin nucleation-promoting factors upstream of Arp2/3 that are expressed in B-cells (Padrick and Rosen 2010). To determine if either or both were responsible for activating Arp2/3 for B-cell contraction, we utilized the N-WASP inhibitor wiskostatin (Wisko), and splenic B-cells from B-cell-specific N-WASP knockout mice (cNKO) and germline WASP knockout (WKO) mice (Westerberg et al. 2012; Liu et al. 2013a). Wisko has been shown to inhibit N-WASP activation while enhancing WASP activation in B-cells (Figure 1-figure supplement 4) (Liu et al. 2013a). We found that both Wisko (10 µM) (Figure 1D and E and Figure 1-Video 1B) and cNKO (Figure 1F and G and Figure 1-Video 1C), but not WKO (Figure 1H and I and Figure 1-Video 1D), significantly reduced the percentage of B-cells undergoing contraction, compared to the vehicle, flox, or WT controls. These results suggest that N-WASP- but not WASP-activated Arp2/3 mediates branched actin polymerization for B-cell contraction.

Arp2/3, downstream of N-WASP, generates inner F-actin foci, driving B-cell contraction

To understand how Arp2/3 drives B-cells to transition from spreading to contraction, we identified F-actin structures associated with contracting B-cells that were sensitive to CK-666 treatment. We visualized F-actin by phalloidin staining and compared F-actin organization in the contact zone of B-cells at the spreading (2 min) and contraction (4 min) phases using TIRF. While B-cells in both spreading and contraction phases exhibited phalloidin staining outlining the contact zone (Fig. 2A, green arrows), only B-cells in the contracting phase showed interior phalloidin patches brighter than the phalloidin staining at the periphery of the contact zone (Figure 2A, purple arrows). These F-actin patches were organized into a ring-like structure and resided ∼1 μm behind the spreading front, surrounding an F-actin-poor center (Figure 2A, purple arrows). Here, we refer these F-actin patches as inner F-actin foci. We identified inner F-actin foci based on whether their peak fluorescence intensity (FI) was ≥2 fold of the mean fluorescence intensity (MFI) of phalloidin in the no-foci area, had diameters of ≥250 nm, and were located 1 µm away from the edge of the contact zone. We found that such inner F-actin foci were detected in >60% of B-cells in the contracting phase (4 min) but only in <20% of B-cells in the spreading phase (2 min) (Figure 2B). CK-666 treatment at 0 min, which inhibited B-cell contraction, significantly reduced the percentage of B-cells showing inner F-actin foci at 4 min but not at 2 min (Figure 2B). CK-666 treatment did not affect the phalloidin staining outlining the contact zone (Figure 2A). Similarly, the percentage of cNKO B-cells showing inner F-actin foci was drastically reduced at 4 min but not at 2 min (Figure 2C). These results suggest that the formation of these inner F-actin foci is associated with B-cell contraction.

Figure 2.

We further quantified the number of inner F-actin foci in individual B-cells from mice expressing the LifeAct-GFP transgene (Riedl et al. 2010) (Figure 2D, E, and G), which allowed us to monitor F-actin reorganization using live-cell imaging (Figure 2-Video 1), or using phalloidin staining in flox control and cNKO B-cells (Figure 2F). Consistent with the results with phalloidin staining, CK-666 treatment at time 0 significantly reduced the number of inner F-actin foci in the contact zone (Figure 2D and Figure 1-Video 1A and B). When we followed the same B-cells before and after CK-666 treatment at 2 min, many of the inner F-actin foci formed before the treatment (1 min 50 sec) disappeared after the treatment (2 min 20 sec), significantly reducing the number of inner F-actin foci (Figure 2E). Similar to the CK-666 treatment, cNKO cells significantly reduced the number of inner F-actin foci in the contact zone, compared to flox controls (Figure 2F). In contrast, WKO, which does not affect B-cell contraction, did not significantly change the number of inner F-actin foci, compared to WT control B-cells (Figure 2G and Figure 2-Video 1C and D). Thus, N-WASP- but not WASP-activated Arp2/3 drives the formation and the maintenance of contraction-associated inner F-actin foci.

Inner F-actin foci are derived from lamellipodial actin networks supporting the spreading membrane

We next examined the formation of contraction-associated inner F-actin foci utilizing live-cell TIRF imaging of B-cells from mice expressing LifeAct-GFP. We generated kymographs along eight lines from the center of each contact zone using time-lapse images of LifeAct-GFP (Figure 3A and B and Figure 3-figure supplement 1). Analysis of these kymographs showed that most F-actin foci were first detected closely behind lamellipodial F-actin networks. Following maximal spreading, these F-actin foci moved centripetally while increasing in intensity, becoming inner F-actin foci, in cells transitioning from spreading to contraction (Figure 3B, top panel, and Figure 3-figure supplement 1). However, such centripetally moving F-actin foci were not detected in B-cells that did not undergo contraction (Figure 3B, bottom panel). We quantified the percentage of the eight kymographs from each cell that exhibited lamellipodia-derived inner F-actin foci and found, on average, that six out of eight kymographs from contracting cells showed lamellipodia-derived inner F-actin foci, compared to only one or two kymographs from non-contracting cells (Figure 3C). To examine the temporal relationship between the generation of lamellipodia-derived inner F-actin foci and contraction, we plotted the percentage of kymographs with lamellipodia-derived inner F-actin foci over time with the spreading to contracting transition time set as 0 (Figure 3D). We found that the percentage of kymographs showing lamellipodia-derived inner F-actin foci peaked at almost the same time when the spreading transitioned to contraction (Figure 3D), suggesting a close temporal relationship between the two events. Furthermore, the N-WASP inhibitor Wisko, but not WKO, significantly inhibited the formation of lamellipodia-derived inner F-actin foci (Figure 3E). Thus, inner F-actin foci originate from branched actin driven lamellipodia and form simultaneously with the transition of B-cell spreading to contraction.

Figure 3.

N-WASP-activated Arp2/3 generates inner F-actin foci by sustaining the lifetime and the centripetal movement of lamellipodial F-actin

We examined the mechanism by which N-WASP and Arp2/3 generate inner F-actin foci by measuring their relative lifetime and mobility using kymographs generated from TIRF time-lapse images of B-cells expressing LifeAct-GFP. Inner F-actin foci were identified as described above, and their tracks were manually determined (Figure 4A, black dashed lines). The time window in which an inner F-actin patch could be detected in a kymograph was measured as the relative lifetime, as actin foci could move away from the kymograph line or the TIRF evanescent field (Figure 4A, right panels). The distance each F-actin focus moved during its lifetime was used to calculate its speed (Figure 4A, right panels). Compared to CK-689-treated B-cells, CK-666 treatment at time 0 significantly reduced the relative lifetime and the centripetal speed of inner F-actin foci (Figure 4B-D). After B-cells were treated with CK-666 at 2 min (the average time for B-cells to reach the maximal spreading), the relative lifetime and the centripetal speed of inner F-actin foci were significantly lower than before the treatment in the same cell (Figure 4E-G, line-linked dots). Similarly, the N-WASP inhibitor Wisko significantly reduced the relative lifetime and the centripetal speed of inner F-actin foci (Figure 4H-J); however, WKO had no significant effect (Figure 4K-M). These results show that N-WASP and Arp2/3 mediated branched actin polymerization prolongs the lifetime of lamellipodia-derived F-actin foci and drives them to move inward in the B-cell contact zone.

Figure 4.

Inner F-actin foci, generated by N-WASP-activated Arp2/3, facilitate NMII recruitment and reorganization into ring-like structures

As non-muscle myosin II (NMII) is required for B-cell contraction (Seeley-Fallen et al. 2022), we examined the relationship between the formation of inner F-actin foci and the recruitment and reorganization of NMII in the contact zone using TIRF imaging of B-cells from mice expressing a GFP-NMIIA transgene. Upon interacting with Fab’-PLB, the GFP-NMIIA MFI in the contact zone of untreated WT B-cells increased rapidly in the first minute and slowly afterward (Figure 5A and B and Figure 5-Video 1). Wiskostatin treatment significantly reduced the GFP-NMIIA MFI in the contact zone (Figure 5B) and its initial rate of increase (as determined by the slope of the GFP-NMIIA MFI increase at 0∼30 sec time window) (Figure 5C). Kymographs generated from time-lapse TIRF images of B-cells from mice expressing GFP-NMIIA and LifeAct-RFP showed that recruited NMIIA accumulated between lamellipodia and inner F-actin foci when the foci moved centripetally away from lamellipodia (Figure 5D, white arrow, and Figure 5-Video 1). Recruited NMIIA reorganized with inner F-actin foci to form a ring-like structure in the contact zone (Figure 5A, D, and E, and Figure 5-Video 1). The percentage of B-cells with NMIIA ring-like structures, visualized by immunostaining, increased over time as more B-cells underwent contraction (Figure 5E and F). Compared to flox controls, the percentage of cNKO B-cells with NMIIA ring-like structures was significantly decreased (Figure 5E and F). Wiskostatin treatment also reduced NMIIA recruitment and ring-like structure formation (Figure 5-Video 1). Surprisingly, the percentage of WKO B-cells with NMIIA ring-like structure was higher than that of flox control B-cells (Figure 5E and F). Thus, N-WASP and Arp2/3 mediated branched actin polymerization promotes the recruitment and the reorganization of NMII ring-like structures by generating inner F-actin foci in the contact zone.

Figure 5.

B-cell contraction increases the BCR molecular density in individual clusters

To understand how B-cell contraction promotes BCR signaling attenuation, we examined the impact of B-cell contraction on the properties of BCR clusters. We first measured the MFI of AF546-Fab’ attached to PLB gathered by B-cells into the contact zone as an indication of the overall BCR molecular density. The clustering of AF546-Fab’ on PLB by B-cell binding reflects surface BCR clustering, as B-cell binding to transferrin (Tf)-tethered PLB does not cause surface BCRs to cluster and be phosphorylated (Figure 6-figure supplement 1) (Liu et al. 2011; Liu et al. 2013a). The MFI of AF546-Fab’ in the B-cell contact zone increased over time.

Figure 6.

Treatment with CK-666, Wisko, or cNKO all reduced the AF546-Fab’ MFI, particularly during the time window of B-cell contraction in controls (Figure 6A-G, purple rectangles, and Figure 6-Video 1). Notably, the rates of increase in AF546-Fab’ MFI, calculated from the slopes of AF546-Fab’ MFI versus time plots in individual cells, were significantly higher during B-cell contraction than before B-cell contraction in control cells and conditions (Figure 6D-K). Significantly, inhibiting B-cell contraction by CK-666 treatment at 0 (Figure 6A, D, and H, and Figure 6-Video 1A and D) or 2 min (Figure 6A, E, and I), Wisko (Figure 6B, F, and J, and Figure 6-Video 1B and E), and cNKO (Figure 6C, G, and K, and Figure 6-Video 1C and F) abolished the increases in AF546-Fab’ accumulation rates. We further examined the peak FI of AF546-Fab’ in individual microclusters as a measure of the BCR molecular density in individual clusters. AF546-Fab’ clusters were identified based on their diameters ≥250 nm, peak FI ≥1.1 fold outside the B-cell contact zone, and trackable for ≥20 sec (Figure 6L and Figure 6-figure supplement 2). AF546-Fab’ microclusters could not be identified during the early stage of B-cell spreading. Time-lapse imaging by TIRF enabled us to measure the rate of increase in AF546-Fab’ peak FI in individual clusters (Figure 6M and Figure 6-figure supplement 2). Consistent with our observation of AF546-Fab’ MFI increase in the contact zone, the peak FI of individual clusters increased at a faster rate during contraction than after contraction (when the contact area no longer decreased) (Figure 6N). Furthermore, CK-666 (Figure 6O), Wisko (Figure 6P), and cNKO (Figure 6Q) all significantly reduced the rate of increase in AF546-Fab’ peak FI in individual clusters. These results show that B-cell contraction significantly increases the molecular density of BCRs in BCR clusters.

Increased BCR molecular density by B-cell contraction reduces BCR phosphorylation levels in individual microclusters

To examine how BCR molecular density influenced BCR signaling capability, we immunostained B-cells interacting with AF546-Fab’-PLB for 1, 3, 5, and 7 min, for phosphorylated CD79a (pCD79a, Y182) and performed IRM and TIRF imaging (Figure 7A and H). We analyzed equal numbers of AF546-Fab’ clusters selected randomly, in the contact zone of B-cells interacting with Fab’-PLB for 1, 3, 5, and 7 min, using a gradient threshold of their AF546-Fab’ MFI, 1.1 to 4.1 fold of the background, and with diameter ≥250 nm (Figure 6-figure supplement 2). We determined the MFI of pCD79a and AF546-Fab’ in individual clusters and plotted the MFI ratio of pCD79a relative to AF546-Fab’ (reflecting the relative level of BCR phosphorylation) versus the AF546-Fab’ peak FI (reflecting BCR molecular density) in individual clusters (Figure 7B and C). The dot plots show that when the AF546-Fab’ peak FI of individual clusters was relatively low, the MFI ratios of pCD79a to AF546-Fab’ increased with the AF546-Fab’ peak FI in individual clusters of flox control B-cells (Figure 7B). When the AF546-Fab’ peak FI reached a certain level, the MFI ratios of pCD79a to AF546-Fab’ decreased as the AF546-Fab’ peak FI in individual clusters further increased (Figure 7A and B). Inhibition of B-cell contraction by cNKO reduced the AF546-Fab’ peak FI in individual clusters, maintaining it within a relatively low range, where the MFI ratios of pCD79a to AF546-Fab’ increased with the AF546-Fab’ peak FI (Figure 7C and D). Additionally, the average pCD79a to Fab’ MFI ratios were much higher in cNKO B-cells than flox control B-cells, when comparing AF546-Fab’ clusters with the same range of peak FI (Figure 7B-D). During the contraction stage (5 and 7 min), flox control B-cells exhibited increased AF546-Fab’ MFI (Figure 7E) but decreased pCD79a MFI (Figure 7F) and the MFI ratios of pCD79a relative to AF54-Fab’ (Figure 7G) in individual clusters, compared to the spreading stage (1 and 3 min). In contrast, non-contracting cNKO B-cells only slightly increased AF546-Fab’ MFI (Figure 7E) but significantly increased pCD79a MFI (Figure 7F) and the pCD79a to AF546-Fab’ MFI ratios (Figure 7G) in individual clusters at the contraction stage compared to the spreading stage. Consequently, individual clusters in non-contracting cNKO B-cells had significantly lower AF546-Fab’ MFI but significantly higher pCD79a MFI and pCD79a to AF546-Fab’ FIRs than contracting flox control B-cells at 5 and 7 min but not at 1 and 3 min (Figure 7E-G). Similarly, inhibiting B-cell contraction by treatment with the Arp2/3 inhibitor CK-666 at 2 min post-stimulation reduced AF546-Fab’ MFI (Figure 7I) but increased pCD79a MFI (Figure 7J) and the MFI ratios of pCD79a relative to AF546-Fab’ (Figure 7K) in individual clusters. These data suggest that increases in BCR molecular density in BCR clusters during B-cell contraction inhibit BCR phosphorylation.

Figure 7.

Increased BCR molecular density by B-cell contraction promotes the disassociation of the stimulatory kinase Syk from BCR microclusters

Increased BCR molecular density may promote signaling attenuation by inducing the disassociation and/or dephosphorylation of stimulatory kinases from and at BCR clusters. To test this hypothesis, we analyzed the relative amounts of Syk, a major stimulatory kinase in the BCR signaling pathway, and its phosphorylated form pSyk (Y519/520) in individual BCR clusters in relation to the molecular density of BCRs. Splenic B-cells were incubated with Fab’-PLB for 3 and 7 min (when most cells were at the spreading and contraction phase, respectively), fixed, permeabilized, stained for Syk or pSyk, and imaged using TIRF. We measured MFI ratios of Syk relative to AF546-Fab’ in individual BCR clusters to reflect the relative amount of Syk associated with individual BCR clusters and analyzed their relationship with AF546-Fab’ peak FI (reflecting the molecular density within BCR clusters). AF546-Fab’ clusters were detected and analyzed as described above. In flox control B-cells, the highest fractions of BCR clusters had an AF546-Fab’ peak FI at the150∼200 range (Figure 8A, brown line and symbols). The average MFI ratios of Syk to Fab’ in individual clusters increased at a low peak FI range (<140) and did not significantly decrease until Fab’ peak FI reached a relatively high range (>280) (Figure 8A, black line and symbol). B-cells from cNKO mice exhibited a reduction in the Fab’ peak FI of clusters (Figure 8B and C, brown line and symbol) but an increase in the average Syk to Fab’ MFI ratios in clusters in a wide range of Fab’ peak FI, when compared to flox control B-cells (Figure 8B and C, black lines and symbols). However, the average Syk to Fab’ MFI ratios of cNKO B-cells decreased in clusters with Fab’ peak FI ≥220 and reduced to levels similar to those in flox control B-cells in clusters with Fab’ peak FI ≥240 (Figure 8B and C, black lines and symbols). These data suggest that increases in the molecular density of BCR clusters, reflected by Fab’ peak FI, induce disassociation of Syk from BCR clusters in both flox control and cNKO B-cells. Our results also show that BCR clusters in cNKO B-cells have significantly higher levels of Syk association than in flox control B-cells, even though they have similar Fab’ peak intensities.

Figure 8.

We next analyzed the relationship of the pSyk level in individual clusters with BCR molecular density using the method described above. We found that the MFI ratios of pSyk relative to AF546-Fab’ gradually decreased with increases in Fab’ peak FI in both flox control and cNKO B-cells (Figure 8D and E, black lines and symbols), even though we did not observe an increase in pSyk to Fab’ MFI ratio at the low Fab’ peak FI range. Similar to the Syk to Fab’ MFI ratio, the average pSyk to Fab’ MFI ratios of individual clusters in cNKO B-cells were much higher than those in flox control B-cells, except for those at the high Fab’ peak FI range (Figure 8D-F). To confirm this result, we analyzed equal numbers of pSyk clusters in the same cells, based on the criteria of ≥1.3 fold increase in the peak FI compared to the background outside the contact zone with a diameter of ≥250 nm. Similar to clusters identified by AF546-Fab’, the average MFI ratios of pSyk to Fab’ in these pSyk clusters decreased with increase in their Fab’ peak FI in both flox control and cNKO B-cells (Figure 8G-I, black lines and symbols). Again, the average pSyk to Fab’ MFI ratios of pSyk clusters were much higher in cNKO than flox control B-cells, but were reduced to similar levels in clusters with relatively high Fab’ peak FI (Figure 8I, black line and symbols). Similar reductions of total Syk and pSyk with increasing molecular density of BCR clusters suggest that the disassociation of Syk from BCR clusters, caused by cell contraction-induced increases in molecular density, contributes to BCR signaling attenuation.

Increased BCR molecular density by B-cell contraction promotes disassociation of the inhibitory phosphatase SHIP-1 from BCR microclusters

The inhibitory phosphatase SHIP-1 is essential for B-cell signaling attenuation (Brauweiler et al. 2000; Liu et al. 2011), suggesting that increases in the molecular density of BCR clusters by B-cell contraction may promote SHIP-1 recruitment. We used the methods described above to address this hypothesis, staining cells for total SHIP-1 and phosphorylated SHIP-1 (pSHIP-1 Y1020). We found that the average SHIP-1 to Fab’ MFI ratios in both flox control and cNKO B-cells decreased with Fab’ peak FI at similar rates (Figure 9A-C, black lines and symbols), even though inhibition of contraction by cNKO reduced the Fab’ peak FI of BCR clusters (Figure 9A-C, brown lines and symbols). Notably, the reduction in the SHIP-1 to Fab’ MFI ratios with increasing Fab’ peak FI occurred at the lowest detectable Fab’ peak PI (Figure 9A-C, brown lines and symbols), when the Syk to Fab’ MFI ratios increased and were sustained (Figure 8A-C). Furthermore, this reciprocal relationship between the SHIP-1 to Fab’ MFI ratio and Fab’ peak FI continued over the entire Fab’ peak FI range. It also remained the same in both flox control and cNKO B-cells (Figure 9A-C). These results suggest that SHIP-1 disassociates from BCR clusters as their molecular density increases, and that the SHIP-1 disassociation is more sensitive to the molecular density of BCR clusters than Syk disassociation.

Figure 9.

Similar to the relationship of total SHIP-1 with BCR molecular density, the average MFI ratios of pSHIP-1 relative to AF546-Fab’ in individual Fab’ clusters decreased with increases in Fab’ peak FI at similar rates in flox control and cNKO B-cells (Figure 9D-F, black lines and symbols), even though inhibition of contraction by cNKO reduced Fab’ peak FI of BCR clusters (Figure 9D-F, brown lines and symbols). The average pSHIP-1 to Fab’ MFI ratios in individual pSHIP-1 clusters, detected and analyzed in the same way as pSyk clusters, showed the same decrease with increases in their Fab’ peak FI in both flox control and cNKO B-cells (Figure 9G-I, black lines and symbols). Notably, the average pSHIP-1 to Fab’ MFI ratios in individual Fab’ or pSHIP-1 clusters in flox control and cNKO B-cells were at similar levels at the same Fab’ peak FI ranges (Figure 9F and I, black lines and symbols). These results indicate that contraction-induced molecular density increases within individual BCR clusters do not induce preferential recruitment of SHIP-1; instead, it promotes the disassociation of SHIP-1 from BCR clusters.

Discussion

When binding membrane-associated antigen, B-cells undergo actin-mediated spreading followed by a contraction, which amplifies BCR signaling and promotes immunological synapse formation (Fleire et al. 2006; Harwood and Batista 2011). We previously showed that B-cell contraction after spreading on antigen-presenting surfaces promotes BCR signaling attenuation (Liu et al. 2013a; Seeley-Fallen et al. 2022). However, how the actin cytoskeleton reorganizes as B-cells transition from spreading to contraction and how B-cell contraction downregulates BCR signaling have been elusive. Here we demonstrate that inner F-actin foci formed at the contact zone distal to the lamellipodial F-actin network promote B-cells to switch from spreading to contraction. These inner foci are derived from the lamellipodial F-actin network that mediates spreading, are generated by N-WASP-but not WASP-activated Arp2/3-mediated branched actin polymerization, and facilitate NMII recruitment, enabling B-cell contraction. B-cell contraction increases BCR molecular density in existing clusters, which promotes the disassociation of both the stimulatory kinase Syk and the inhibitory phosphatase SHIP-1, leading to signaling attenuation.

One significant finding of this study is that Arp2/3-mediated polymerization of branched actin is required for B-cell contraction. Arp2/3-mediated branched actin polymerization is known to drive B-cell spreading and to create actin centripetal flow at the contact zone between B-cells and antigen-presenting surface (Bolger-Munro et al. 2019). The actin structure that supports B- and T-cell spreading to form the immunological synapse with antigen-presenting cells is similar to lamellipodial F-actin networks found in adherent cells (Bunnell et al. 2001; Koestler et al. 2008; Bolger-Munro et al. 2019). Lamellipodial F-actin networks consist primarily of branched actin filaments polymerizing against the plasma membrane interspersed with bundled actin filaments (Krause and Gautreau 2014; Skau and Waterman 2015). Contractile actin structures, such as stress fibers, are typically generated from bundled actin filaments, as observed in adherent and migrating cells (Levayer and Lecuit 2012; Tojkander et al. 2015; Hammer et al. 2019). Here, we show that inner F-actin foci, generated by Arp2/3-mediated actin polymerization, transition B-cells from spreading to contraction. Furthermore, these inner F-actin foci are directly derived from lamellipodial F-actin networks. Based on the observed dynamics, we infer that instead of polymerizing against the plasma membrane, Arp2/3 appears to nucleate actin polymerization in the opposite direction, sustaining actin foci and their movement away from the lamellipodia. However, this study does not exclude the involvement of bundled actin filaments. Based on the requirement of formin-activated bundled actin filaments for lamellipodial F-actin networks and B-cell spreading (Wang et al. 2022), we can speculate that bundled actin filaments may play a role in the formation and movement of these inner F-actin foci, as well as in NMII recruitment.

Our work provides new insights into distinct functions of the actin nucleation-promoting factors WASP and N-WASP in controlling cell morphology and signaling. Immune cells, including B-cells, express both hematopoietic-specific WASP and the ubiquitous homolog of WASP, N-WASP. These two share high sequence homology, activation mechanism, and Arp2/3 activation function (Padrick and Rosen 2010). We previously identified distinct functions of these two factors that are unique to B-cells. While both are required for B-cell spreading, N-WASP plays a unique role in B-cell contraction. WASP and N-WASP double knockout B-cells fail to spread on antigen-presenting surfaces, but B-cell-specific N-WASP knockout enhances B-cell spreading and delays B-cell contraction (Liu et al. 2013a). Furthermore, these two factors reach peak activation at different times (phosphorylated WASP peaks during B-cell spreading, while phosphorylated N-WASP peaks during contraction) and are detected at distinct locations (the majority of WASP is at the periphery, and N-WASP is scattered across the B-cell contact zone) (Liu et al. 2013a). Here, we reveal the exact role of N-WASP in B-cell contraction – to generate inner F-actin foci from lamellipodial F-actin networks by activating Arp2/3-mediated actin polymerization. N-WASP- but not WASP-activated actin polymerization prolongs the relative lifetime of actin foci and mediates their inward motion toward the center of the contact zone. The delayed activation time and the location in the interior of the contact zone likely give N-WASP a unique opportunity to generate inner F-actin foci. As WASP and N-WASP suppress each other for activation in B-cells (Liu et al. 2013a), the timing and the location of their activation may depend on the relative levels of their expression and their abilities to compete for upstream and downstream molecules. As cNKO only delays and reduces but does not block B-cell contraction and inner F-actin foci formation, other actin factors may be involved, such as the WASP-family verprolin-homologous protein (WAVE) (Rotty et al. 2013).

Our recently published data show that NMII is required for B-cell contraction. Activated NMII is recruited to the B-cell contact zone in a SHIP-1-dependent manner and forms a peripheral ring surrounding the contact zone during B-cell contraction (Seeley-Fallen et al. 2022). Here, we further show that in addition to SHIP-1, N-WASP but not WASP is also involved in NMII recruitment and NMII ring-like structure formation, probably by generating inner F-actin foci. Collectively, these findings suggest that SHIP-1 coordinates with N-WASP to recruit NMII. Our previous finding that SHIP-1 promotes B-cell contraction by facilitating N-WASP activation (Liu et al. 2011; Liu et al. 2013a) supports this notion. However, it is surprising that the recruitment of NMII activated by BCR signaling to the contact zone is associated with inhibitory signaling molecules. BCR downstream signaling, including Ca²⁺ fluxes and Rho-family GTPase activation, likely activates NMII motor activity (Vicente-Manzanares et al. 2009). The activation switches NMII from the incompetent folded conformation to the competent extended conformation, enabling NMII molecules to bind to F-actin and assemble into contractible bipolar filaments and stacks (Matsumura 2005). A denser F-actin organization has been shown to increase NMII filament stacking (Fenix and Burnette 2018). Here, we showed that the inner F-actin foci generated by N-WASP-activated Arp2/3 are more stable and denser and thus likely promote NMII binding and stacking more efficiently than F-actin generated by WASP-activated Arp2/3. Indeed, our kymograph analysis revealed similar time windows and spatial locations for NMII accumulation and for generating inner F-actin foci from lamellipodial F-actin, supporting our hypothesis.

Wang et al. (Wang et al. 2022) recently showed that B-cells generate actomyosin arcs when interacting with membrane-associated antigen in the presence of adhesion molecules, as when interacting with professional antigen-presenting cells. In contrast to T-cells, B-cells can also respond to membrane-associated antigen without the help of adhesion molecules, when engaging antigen on pathogenic cells, like bacteria, parasites, and cancer cells. While the relationship between actomyosin arcs and inner F-actin foci remains to be explored, the role of inner F-actin foci in the formation of NMII ring-like structures revealed in our study suggests that the inner F-actin foci in the absence of adhesion molecules may be the facilitator or precursors of the actomyosin arcs. These structures can potentially be enhanced and matured by interactions between B-cells and antigen-presenting membranes mediated by adhesion molecules.

Our findings suggest an increase in the molecular density within BCR clusters or B-cell synapses as one of the mechanisms by which B-cell contraction promotes BCR signaling attenuation. During contraction, the molecular density of individual BCR clusters, measured by their MFI and peak FI of BCR clustered Fab’, increased faster than during the spreading and post-contraction phases. Surprisingly, increases in molecular density induced disassociation of both the stimulatory kinase Syk and the inhibitory phosphatase SHIP-1, contradicting the existing dogma of sequential association of stimulatory and inhibitory signaling molecules with the BCR (Franks and Cambier 2018). Interestingly, the disassociation of SHIP-1 is much more sensitive to BCR molecular density than Syk disassociation. Syk does not disassociate until the molecular density of BCR clusters reaches the top 10% range, while SHIP-1 disassociates from BCR clusters with increases over the entire detectable range of molecular densities. When B-cell contraction is inhibited, the molecular density of BCR clusters is reduced to a range with limited Syk disassociation and normal SHIP-1 disassociation, resulting in high Syk to SHIP-1 molecular ratios in BCR clusters than contracting B-cells, increasing BCR signaling levels. The resolution of our widefield TIRF imaging limits our ability to examine nascent BCR clusters that actively recruit signaling molecules. However, the consistency of the data from AF546-Fab’ and pSyk or pSHIP-1 clusters confirms our results. How the molecular density within BCR clusters promotes disassociation of signaling molecules is unknown. We speculate that molecular crowding may displace signaling molecules out of BCR clusters. Supporting this possibility, disassociation of the 145 kDa SHIP-1 is likely to be more sensitive to changes in the molecular density of BCR clusters, compared to disassociation of the 72 kDa Syk. However, these hypotheses need to be further examined.

An interesting finding of this study is that the pCD79a and pSyk levels were higher in individual BCR clusters of cNKO B-cells than those of flox control B-cells, even when BCR clusters had similar molecular densities. These data suggest additional mechanisms for B-cell contraction to promote BCR signaling attenuation. In addition to increasing molecular density within BCR clusters, the contractile forces generated by actomyosin rings on the B-cell membrane and BCR clusters may cause changes to the conformations of both BCRs and associated signaling molecules and their lateral interactions in the membrane. These changes may favor the disassociation rather than the association of signaling molecules with BCR clusters. Recently solved molecular assembly structures of both human and mouse BCRs (Dong et al. 2022; Su et al. 2022) support conformational and lateral molecular interaction changes as possible mechanisms for BCR signaling regulation. Actomyosin-mediated contractile forces were also shown to be required for BCR-mediated endocytosis of antigen (Natkanski et al. 2013; Hoogeboom et al. 2018; Maeda et al. 2021), which removes the BCR from the cell surface, downregulating the BCR surface signaling. Whether the inner F-actin foci identified here play any role in BCR endocytosis remains an interesting question. Our early finding that N-WASP is required for BCR endocytosis (Liu et al. 2013a) supports this notion. Similar to the differential impact of the molecular density on the disassociation of Syk and SHIP-1 from BCR clusters, B-cell contraction may differentially affect the interaction of surface BCRs with cytoplasmic and membrane-anchored signaling molecules, such as the lipidated and lipid raft-resident Src kinase Lyn that is responsible for phosphorylating both the immunoreceptor tyrosine-based activation motif (ITAM) of CD79a and the immunoreceptor tyrosine-based inhibitory motif (ITIM) that SHIP-1 binds to (Franks and Cambier 2018).

The results presented here have revealed novel insights into the mechanisms underlying actin-facilitated signaling attenuation of the BCR. Taking our previous and current data together, we propose a new working model for such actin-mediated signaling downregulation. Upon maximal spreading, N-WASP distal to lamellipodial networks activates Arp2/3-mediated branched actin polymerization, which sustains actin foci originating from lamellipodia and their movement away from the lamellipodia towards the center of the B-cell contact zone, generating inner F-actin foci. NMII is then preferentially recruited to these relatively stable inner foci, forming ring-like actomyosin structures, which enable B-cell contraction. B-cell contraction pushes the BCR microclusters formed during B-cell spreading to the center of the contact zone, increasing their molecular density. Increased molecular density promotes disassociation of signaling molecules from BCR clusters, probably due to crowding and conformational changes, leading to signal downregulation. Inhibitory signaling molecules likely activate the actin reorganization that drives B-cell contraction, as both the activation of N-WASP and the recruitment of NMII to the B-cell contact zone depend on SHIP-1 (Liu et al. 2013a; Seeley-Fallen et al. 2022). Thus, actin reorganization downstream of inhibitory phosphatases reinforces signaling attenuation by driving B-cell contraction.

Materials and Methods

Key Reagents Table

Mice and B-cell isolation

Wild-type (WT) C57BL/6 mice and WASP knockout (WKO) mice on the C57BL/6 background were purchased from Jackson Laboratories. C56BL/6 mice expressing LifeAct-GFP transgene (Riedl et al. 2010) were kindly provided by Dr. Roberto Weigert at National Cancer Institute, USA. C56BL/6 mice expressing the LifeAct-RFP transgene were kindly provided by Dr. Klaus Ley at La Jolla Institute for Allergy & Immunology. C57BL/6 mice expressing the GFP-non-muscle myosin IIA (NMIIA) transgene were kindly provided by Dr. Robert Adelstein at the National Heart, Lung, and Blood Institute, USA. WKO mice were bred with LifeAct-GFP mice to obtain LifeAct-GFP-expressing WKO mice. GFP-NMIIA mice were crossed with LifeAct-RFP mice to generate mice expressing both transgenes. N-WASP^flox/flox on a 129Sv background (Cotta-de-Almeida et al. 2007) were kindly provided by Dr. Lisa Westerberg at Karolinska Institute, Sweden. B-cell-specific N-WASP knockout mice (cNKO, CD19^Cre/+N-WASP^flox/flox) and floxed littermate controls (N-WASP^flox/flox) were obtained by breeding N-WASP^flox/flox mice with CD19^Cre/+ mice on a C57BL/6 background.

Primary B-cells were isolated from the spleens of 6 to 18 week-old male or female mice, using a previously published protocol (Sharma et al. 2009). Briefly, mononuclear cells were isolated by Ficoll density-gradient centrifugation (Millipore Sigma), and T-cells were eliminated by complement-mediated cytolysis with anti-mouse CD90.2 mAb (BD Biosciences) and guinea pig complement (Innovative Research Inc). Monocytes and dendritic cells were eliminated by panning at 37°C and 5% CO₂. Isolated B-cells were kept on ice in DMEM (Lonza) supplemented with 0.6% BSA (Thermo Fisher Scientific). All work involving mice was approved by the Institutional Animal Care and Usage Committee of the University of Maryland.

Pseudo-antigen-coated planar lipid bilayers

Planar lipid bilayers (PLB) were prepared using a previously described method (Dustin et al. 2007; Liu et al. 2012a). Briefly, liposomes were generated from a mixture of 5 mM (total concentration) 1,2-dioleoyl-sn-glycero-3-phosphocholine and 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine-cap-biotin (Avanti Polar Lipids) at a 100:1 molar ratio by sonication. Glass coverslips, cleaned overnight with Piranha solution (KMG chemicals), were attached to 8-well chambers (Lab-Tek) and incubated with liposomes (0.05 mM) in PBS for 20 min at room temperature and washed with PBS. The chambers were then incubated with 1 µg/ml streptavidin (Jackson ImmunoResearch Laboratories) for 10 min, washed with PBS, and then incubated with 10 µg/ml mono-biotinylated Fab’ fragment of goat anti-mouse Ig(G+M) (pseudo-antigen) (Fab’-PLB) for 10 min, followed by PBS wash. For transferrin-coated PLB (Tf-PLB), 16 16 µg/ml biotinylated holoTF was used. To visualize Fab’ clustering, a mixture of 0.5 µg/ml Alexa Fluor (AF) 546-labeled and 9.5 µg/ml unlabeled mono-biotinylated Fab’ fragment of goat anti-mouse Ig(G+M) was used. The lateral mobility of AF546-Fab’ on the PLB was tested using fluorescence recovery after photobleaching (FRAP) to ensure ≥85% AF546 fluorescence recovery within 1 min after photobleaching using a Zeiss LSM 710 equipped with a 60X oil-immersion objective.

Mono-biotinylated Fab’ fragments were generated as previously described (Liu et al. 2011). Briefly, the disulfide bond linking the two Fab fragments of F(ab’)₂ goat anti-mouse Ig(G+M) was reduced using 2-mercaptoethylamine HCL (Thermo Fisher Scientific) and biotinylated by using maleimide-PEG₂-biotin (Thermo Fisher Scientific). Mono-biotinylated Fab’ fragments of goat anti-mouse Ig(G+M) were labeled with AF546 using an Alexa Fluor 546 antibody labeling kit (Thermo Fisher Scientific). The molar ratio of AF546 to Fab’ in the AF546-labeled Fab’ was ∼2 determined by a Nanodrop spectrophotometer (Nanodrop Technologies).

Total internal fluorescence microscopy

To visualize molecules proximal to interacting sites between B-cells and Fab’- or Tf-PLB, we utilized total internal reflection fluorescence microscopy (TIRF) and interference reflection microscopy (IRM). Images were acquired using a Nikon TIRF system on an inverted microscope (Nikon TE2000-PFS, Nikon Instruments Inc.) equipped with a 60X, NA 1.49 Apochromat TIRF objective (Nikon), a Coolsnap HQ2 CCD camera (Roper Scientific), and two solid-state lasers of wavelength 491 and 561 nm. IRM, AF488, and AF546 images were acquired sequentially.

The plasma membrane area of B-cells contacting PLB was determined using IRM images and custom MATLAB codes. Whether a B-cell contracted or not was determined using the area versus time plots, wherein if a B-cell’s contact zone area reduced by ≥5% for at least 10 sec after reaching a maximum value, it was classified as contracting. To image intracellular molecules, B-cells were incubated with Fab’-PLBs for varying lengths of time in PBS at 37°C, fixed with 4% paraformaldehyde, permeabilized with 0.05% saponin, and stained for various molecules. For live-cell imaging, B-cells were pre-warmed to 37°C and imaged as soon as B-cells were dropped into coverslip chambers coated with Fab’-PLB containing PBS in a humidity chamber at 37°C, at 2 sec per frame, up to 7 min. All images from multiple independent experiments were analyzed using NIH ImageJ and custom MatLab codes. Acquired fluorescence intensity (FI) data were normalized to the one with the lowest FI.

Inhibitors

CK-666 (50 µM, Millipore Sigma) was used to perturb Arp2/3 activity (Nolen et al. 2009), and its non-functional derivative CK-689 (50 µM, Millipore Sigma) as a control. CK-666 or CK-689 was added at either 0 min, the start of incubation with Fab’-PLB, or at 2 min when most B-cells reached maximum spreading. Notably, B-cells take approximately 1 to 1.5 min to land on PLB. The time when B-cells were added to Fab’-PLB is referred to as 0 min, and the time when B-cells landed on PLB as the start of spreading. The effectiveness of CK-666 was determined by its inhibitory effects on the recruitment of Arp2/3, stained by an anti-Arp2 antibody (Abcam), to the B-cell contact zone using TIRF (Figure 1-figure supplement 2). Wiskostatin (Wisko, 10 µM, Millipore Sigma) was used to perturb N-WASP activity in B-cells (Peterson et al. 2004), with DMSO used as a vehicle control. Splenic B-cells were pretreated with Wisko for 10 min at 37°C before and during incubation with Fab’-PLB. The effectiveness of Wisko was determined by its inhibitory effects on the level of phosphorylated N-WASP in the B-cell contact zone using immunostaining and TIRF (Figure 1-figure supplement 4A and C). Possible effects of Wisko on WASP activation were evaluated by measuring the mean fluorescence intensity (MFI) of phosphorylated WASP in the contact zone of B-cells treated with or without Wisko for 10 min (Figure 1-figure supplement 4B and D).

Analysis of the actin cytoskeleton

F-actin was visualized by phalloidin staining in fixed cells and by LifeAct-GFP or LifeAct-RFP expressed by primary B-cells from transgenic mice in live cells. Inner F-actin foci, visualized by both phalloidin staining and LifeAct-GFP, were identified using NIH ImageJ based on the following three criteria: 1) ≥250 nm in diameter, 2) peak fluorescence intensity (FI) ≥2 fold higher than the FI of a nearby area containing no foci, and 3) ≥1 µm away from the outer edge of the B-cell contact zone. The horizontal or vertical length of a focus was measured, and the lesser of the two values was used as its diameter. To analyze the spatiotemporal relationship between inner F-actin foci and the lamellipodial F-actin network, we generated 8 radially and equally spaced kymographs from each cell using time-lapse images and MATLAB. Each kymograph was either classified as contracting or not contracting based on the movement of the leading edge of the B-cell contact zone. The percentage of 8 kymographs from each cell exhibiting inner F-actin foci that could be traced back to lamellipodial F-actin when the B-cell switched from spreading to contraction was determined.

The relative speeds and lifetimes of F-actin foci were determined using three randomly positioned kymographs from each B-cell and F-actin focuses emerging during a 60-sec window right after B-cell maximal spreading. Only those remaining visible in individual kymographs for at least 4 sec were analyzed. The time duration of each F-actin foci detected in a kymograph was used to determine the relative lifetime. The distances individual F-actin foci moved during their lifetimes were used to determine the relative speed. The term ‘relative’ has been used because the disappearance of an F-actin focus could be due to its movement away from the region used for the kymograph or vertically from the TIRF evanescent field of excitation.

To determine the rate of myosin recruitment, GFP-NMIIA MFI in the B-cell contact zone was plotted over time, and the slope for the initial segment of the MFI versus the time plot was determined using linear regression. The percentage of B-cells with the NMIIA ring was determined by visual inspection.

Analysis of BCR-Fab’ clusters

BCR clusters were identified by clustering of AF546-Fab’ on PLBs using custom code by MATLAB. We utilized the median FI of the Fab’-PLB within the same area outside but near the B-cell contact zone as the background. The median, but not mean FI value was used for background FI calculations to minimize fluctuations due to debris. To ensure that individual clusters with varying FI were detected as distinct objects, we used 16 graded thresholds from 1.1 to 4.1 fold of the background (0.2 fold apart) to acquire 16 sets of binary masks for each frame of time-lapse images from each cell. When clusters were detected by multiple thresholds at the same location, only the one identified by the highest threshold was retained. Objects that were smaller than 250 nm in diameter or could not be tracked for at least 20 sec were eliminated. This allowed us to detect the position of BCR-Fab’ clusters and track them over time until they merged with other clusters. The horizontal or vertical length of a focus was measured, and the lesser of the two values was used as its diameter. We chose the peak FI as a metric for the extent of Fab’ clustering by the BCR, because it does not rely on the area occupied by each cluster. The rate of peak FI increase was determined by linear regression of peak FI versus time curves for each cluster for a given window of time.

BCR signaling

Splenic B-cells were incubated with AF546-Fab’-PLB at 37°C and fixed at 1, 3, 5, 7, and 9 min. After fixation, cells were stained for phosphorylated CD79a (pCD79a, Y182) (Cell Signaling Technology), Syk (pSyk, Y519/520) (Cell Signaling Technology), SHIP-1 (pSHIP-1, Y1020) (Cell Signaling Technology), or total Syk or SHIP-1 proteins (Thermo Fisher Scientific). We identified BCR-Fab’ clusters as described. We determined the peak FI of AF546-Fab’ and the MFI of AF546-Fab’, pCD79a, pSyk, pSHIP-1, Syk, or SHIP-1 within each cluster. The ratio of pCD79a MFI to AF546-Fab’ MFI in the same cluster was used to estimate the relative phosphorylation level of BCRs in that cluster. This MFI ratio was plotted against AF546-Fab’ peak FI to analyze the relationship between BCR phosphorylation and BCR density of individual clusters. We calculated the MFI ratio of Syk to Fab’ or SHIP-1 to Fab’ in individual clusters to estimate the relative recruitment level of Syk or SHIP-1 to BCR clusters. We plotted these MFI ratios against Fab’ peak FI to determine the relationship between Syk and SHIP-1 recruitment levels and BCR density in individual clusters. We calculated the MFI ratio of pSyk to Fab’ or pSHIP-1 to Fab’ to estimate the amount of pSyk or pSHIP-1 relative to BCR in individual clusters. We plotted these MFI ratios against Fab’ peak FI to determine the relationship between Syk or SHIP-1 phosphorylation and BCR density in individual clusters. We identified pSyk or pSHIP-1 puncta using the criteria: ≥1.3 fold of the background staining outside the B-cell contact zone and ≥250 nm in diameter. We calculated the MFI ratio of pSyk to Fab’ or pSHIP-1 to Fab’ in individual puncta to estimate the amount of pSyk or pSHIP-1 relative to BCR in individual puncta. We plotted these MFI ratios against Fab’ peak FI to determine the relationship between Syk or SHIP-1 phosphorylation and BCR density in individual puncta.

Statistical analysis

Statistical analysis was performed using the Mann-Whitney U non-parametric test for unpaired groups having different sample sizes or the Student’s t-test for paired groups with the same sample size. To compare curves, Kolmogorov-Smirnov test was used. Statistical analyses were performed in Microsoft Excel, GraphPad Prism, and MATLAB. All data are presented as mean ± SEM (standard error of the mean). When testing multiple hypotheses, p--values acquired using t-tests were corrected using the Benjamini-Hochberg/Yekutieli method for false discovery rate control.

MATLAB scripts

All MATLAB scripts used for this study are available as supplemental materials.

Acknowledgements

This work was supported by NIH grants GM064625 (NWA and WS), GM145313 (AU), and AI122205 (AU and WS). We would like to acknowledge Amy Beaven and Kenneth Class for technical support from the Confocal Microcopy and Flow Cytometry cores. We would like to thank Dr. Norma W. Andrews (University of Maryland, College Park) for her careful reading and critical comments on the manuscript.